Background: In recent decades, two lines of research, phenomenographic and meta-cognitive, have examined students' approaches and epistemological beliefs about learning. To date there has been very little research describing the change in epistemological beliefs in European secondary students, or analysing interrelationships between epistemological beliefs and approaches in order to explain their influence on academic performance. Aims: The first aim of this investigation is to analyse the change in epistemological beliefs and learning approaches in secondary students as they progress through their studies. The second aim is to examine the effects of epistemological beliefs on learning approaches, and learning approaches on academic performance. Sample: About 1,600 Spanish students, boys and girls, from several secondary schools took part in the study. They were between 12 and 20 years old and their average age was 14.79 years. Methods: Measures of epistemological beliefs (EQ: Epistemological Questionnaire), learning approaches (LPQ: Learning Process Questionnaire), and academic performance were obtained. Confirmatory factor analysis was used to examine the dimensionality of the EQ and LPQ questionnaires. In order to achieve our two aims, different statistical techniques were used: MANOVA and ANOVA for our first aim, and structural equation modelling for our second aim. Results: Throughout secondary education epistemological beliefs undergo change, becoming more realistic and complex, and deep-approach scores decline significantly. It was shown that, as predicted, epistemological beliefs influenced academic achievement directly, and also indirectly via students' learning approaches. Conclusions: Our findings point to two conclusions. First, epistemological beliefs and learning approaches change as pupils advance in their studies. Second, the relationship between epistemological beliefs and academic achievement is mediated by approaches to learning.